The broad goal of this proposal is to further develop the candidate's skills as a health services researcher in the area of patient treatment preferences, cost and miscarriage management. The proposed program uses didactic course work, institutional resources and strong mentorship to provide training in claims analysis, survey methodology and cost-effectiveness analysis. The purpose of this proposal is to understand how early pregnancy failure (EPF) is currently managed and to examine the effect of using patient preferences to [unreadable] determine treatment on cost. Miscarriage management provides an ideal opportunity to study and encourage patient-centered care because it is a common event and there are several successful treatment options. Traditionally, management of pregnancy loss has been dilation and curettage (D&C), most often occurring in an operating room. This practice was established during a time when patients with EPF typically presented with acute hemorrhage and/or infection. Today, technological advances allow a non-viable pregnancy to be diagnosed well before the onset of these symptoms. Under these conditions, treatment options include expectant management and drug therapy to cause the uterus to expel the non-viable products. Alternatively, surgical management can be done in an office-based setting, which offers significant cost savings over the same procedure performed in an operating room. The few studies available suggest treatment patterns do not reflect contemporary clinical presentations, patient preference, cost burden or clinical outcomes. We propose to quantitatively describe treatment patterns, or usual care, for EPF management by using analysis of claims data. (Specific Aim I) Our preliminary data presented in this proposal demonstrate that patient treatment preferences may not support routine operative room management schemes. There are no data describing provider factors associated with treatment patterns or choices. Specific Aim II uses population based surveys to examine provider and patient treatment preferences and factors associated with treatment choices. Finally, we use our findings Aims I and II to construct a hypothetical care model that includes all treatment options. Specific Aim III compares the cost of this hypothetical care model which is driven by patient preferences with usual care defined in Aim I. [unreadable] [unreadable] [unreadable]